fix: fetch openai quick start models

Use OpenAI's SDK default base URL for Quick Start model discovery without writing that default into the saved provider config.

nanobot/cli/onboard.py

@@ -65,6 +65,9 @@ _QUICK_START_PROVIDER_KEYS = (
     "zhipu",
 )
 _QUICK_START_CUSTOM_PROVIDER_CHOICE = "Other OpenAI-compatible"
+_QUICK_START_MODEL_FETCH_API_BASES = {
+    "openai": "https://api.openai.com/v1",
+}
 
 _QUICK_START_STEPS = ("Provider + key", "WebUI", "Review")
 
@@ -1513,8 +1516,11 @@ def _configure_quick_start_provider(config: Config) -> bool:
         provider_config.api_base = api_base
 
     model = None
-    if provider_config.api_base:
-        model = _fetch_first_quick_start_model(provider_config.api_base, api_key)
+    model_api_base = provider_config.api_base or _QUICK_START_MODEL_FETCH_API_BASES.get(
+        provider_name
+    )
+    if model_api_base:
+        model = _fetch_first_quick_start_model(model_api_base, api_key)
     if not model:
         model = _input_model_with_autocomplete("Model ID", "", provider_name)
     if not model or not model.strip():

tests/agent/test_onboard_logic.py

@@ -967,6 +967,33 @@ class TestMainMenuUpdate:
         assert config.model_presets["primary"].provider == "openrouter"
         assert config.model_presets["primary"].model == "openai/gpt-4o-mini"
 
+    def test_quick_start_openai_fetches_models_without_storing_base(self, monkeypatch):
+        """OpenAI should still support key-only setup via its SDK default endpoint."""
+        config = Config()
+        calls: dict[str, str] = {}
+
+        monkeypatch.setattr(onboard_wizard, "_show_quick_start_progress", lambda *_args: None)
+        monkeypatch.setattr(onboard_wizard, "_select_with_back", lambda *a, **kw: "OpenAI")
+        monkeypatch.setattr(onboard_wizard, "_input_text", lambda *a, **kw: "sk-openai-test")
+
+        def fake_fetch(api_base, api_key):
+            calls["api_base"] = api_base
+            calls["api_key"] = api_key
+            return "gpt-4o-mini"
+
+        monkeypatch.setattr(onboard_wizard, "_fetch_first_quick_start_model", fake_fetch)
+
+        assert onboard_wizard._configure_quick_start_provider(config) is True
+
+        assert calls == {
+            "api_base": "https://api.openai.com/v1",
+            "api_key": "sk-openai-test",
+        }
+        assert config.providers.openai.api_key == "sk-openai-test"
+        assert config.providers.openai.api_base is None
+        assert config.model_presets["primary"].provider == "openai"
+        assert config.model_presets["primary"].model == "gpt-4o-mini"
+
     def test_quick_start_custom_base_url_fetches_first_model(self, monkeypatch):
         """Unknown providers should use only the user-provided base URL to fetch models."""
         config = Config()